Estimators for the difference in means/medians (2 related samples)
Estimators for the difference in means/medians of related samples and their properties and assumptions.
| Estimator | Purpose |
|---|---|
| Mean difference | Estimate the difference between the means. |
| Standardized mean difference | Estimate the standardized difference between the means. Cohen's d is the most popular estimator using the difference between the means divided by the pooled sample standard deviation. Cohen's d is a biased estimator of the population standardized mean difference (although the bias is very small and disappears for moderate to large samples) whereas Hedge's g applies an unbiasing constant to correct for the bias. |
| Median difference | Estimate the median of the differences. |
| Hodges-Lehmann location shift | Estimate the shift in location. A shift in location is equivalent to a difference between means/medians when the distribution of the differences is symmetric. Symmetry is quite often inherently satisfied for paired data. |
Related concepts
Related tasks
Available in Analyse-it Editions
Standard edition
Method Validation edition
Quality Control & Improvement edition
Ultimate edition
Standard edition
Method Validation edition
Quality Control & Improvement edition
Ultimate edition